-
Notifications
You must be signed in to change notification settings - Fork 0
/
SSA_my_onlyweight.py
173 lines (146 loc) · 7.09 KB
/
SSA_my_onlyweight.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
import math
import numpy as np
import matplotlib.pyplot as plt
from tqdm import tqdm
def initialization_chaotic_map(population, dim):
x = np.zeros((population, dim))
x[0] = np.random.rand(dim)
for i in range(1, population):
a = np.random.rand()
temp_x = np.zeros(dim)
for j in range(dim):
temp_x[j] = math.sin((a * math.pi) / x[i - 1][j])
x[i] = temp_x
return x
def my_sparrow_search_optimization(population, max_iterations, search_num_l, search_num_u, dim, fitness_function, flag_p):
weight_min = 0
weight_max = 2
ST = 0.8
propotion_alerter = 0.1
# The propotion of producer
propotion_producer = 0.2
producer_num = round(population * propotion_producer)
low_bundary = search_num_l * np.ones((1, dim))
up_bundary = search_num_u * np.ones((1, dim))
# 初始化麻雀位置和适应度值
position = np.zeros((population, dim))
fitness = np.zeros(population)
R2_new = np.zeros(population)
chaotic_value = initialization_chaotic_map(population, dim)
for i in range(population):
position[i, :] = low_bundary + (up_bundary - low_bundary) * chaotic_value[i]
fitness[i] = fitness_function(position[i, :])
# 初始化收敛曲线
convergence_curve = np.zeros(max_iterations)
new_p, new_f = None, None
# for t in tqdm(range(max_iterations), desc="SSA_my", miniters=max_iterations/5):
for t in range(max_iterations):
# 对麻雀的适应度值进行排序，并取出下标
fitness_sorted_index = np.argsort(fitness.T)
best_finess = np.min(fitness)
best_finess_index = np.argmin(fitness)
best_position = position[best_finess_index, :]
worst_fitness = np.max(fitness)
worst_fitness_index = np.argmax(fitness)
worst_positon = position[worst_fitness_index, :]
# 计算每个个体的惯性权重
# ———————————反向惯性权重
inertia_weight = np.zeros(population)
inverted_weight = np.zeros(population)
for i in range(population):
inertia_weight[i] = weight_min + (weight_max - weight_min) / 2 * (math.sin(
math.pi * ((fitness[i] - best_finess) / (worst_fitness - best_finess)) + math.pi/2) + 1)
inverted_weight[i] = (weight_min + weight_max) - inertia_weight[i]
# 1) 发现者（探索者、生产者）位置更新策略
# R2 = np.random.rand(1)
for i in range(producer_num):
R2 = np.random.rand(1)
p_i = fitness_sorted_index[i]
if R2 < ST:
alaph = np.random.rand()
new_p = position[p_i, :] * np.exp(-i / (alaph * max_iterations))
# position[p_i, :] = position[p_i, :] * np.exp(-i / (alaph * max_iterations))
elif R2 >= ST:
q = np.random.normal(0, 1, 1)
l_dim = np.ones(dim)
cauchy_gauss = 1 + R2 * np.random.standard_cauchy() + (1 - R2) * np.random.normal(0, 1)
new_p = inertia_weight[p_i] * position[p_i, :] + q * l_dim
# 越界处理
new_p = np.clip(new_p, search_num_l, search_num_u)
new_f = fitness_function(new_p)
# if new_f < fitness[p_i]:
position[p_i, :] = new_p
fitness[p_i] = new_f
# 找出最优的”探索者“
next_best_position_index = np.argmin(fitness[:])
next_best_position = position[next_best_position_index, :]
# 2) 追随者(scrounger)位置更新策略
for i in range(0, population - producer_num):
s_i = fitness_sorted_index[i + producer_num]
o_i = i + producer_num
if o_i > (population / 2):
q = np.random.normal(0, 1 , 1)
r = np.random.uniform(search_num_l, search_num_u, 1)
new_p = q * np.exp((worst_positon - position[s_i, :]) / (o_i ** 2))
# new_p = inverted_weight[s_i] * q * np.exp((worst_positon - position[s_i, :])/(o_i**2))
# new_p = r * np.exp((worst_positon - position[s_i, :]) / (o_i ** 2))
# new_p = (inertia_weight[s_i] * best_position *
# q * np.exp((worst_positon - position[s_i, :])/(o_i**2)))
else:
l_dim = np.ones(dim)
a = np.floor(np.random.rand(1, dim) * 2) * 2 - 1
a_plus = 1 / (a.T * np.dot(a, a.T))
new_p = (next_best_position +
inverted_weight[s_i] * l_dim * np.dot(np.abs(position[s_i, :] - next_best_position), a_plus))
# 越界处理
new_p = np.clip(new_p, search_num_l, search_num_u)
new_f = fitness_function(new_p)
# if new_f.all() < fitness[s_i].all():
position[s_i, :] = new_p
fitness[s_i] = new_f
# 3) 意识到危险的麻雀的位置更新
arrc = np.arange(len(fitness_sorted_index[:]))
random_arrc = np.random.permutation(arrc) # 随机排列序列
num_alerter = round(propotion_alerter * population) # 随机选取警戒者
alerter_index = fitness_sorted_index[random_arrc[0:num_alerter]]
for i in range(num_alerter):
a_i = alerter_index[i]
f_i = fitness[a_i]
f_g = best_finess
f_w = worst_fitness
if f_i > f_g:
beta = np.random.normal(0, 1 , 1)
new_p = (best_position +
inverted_weight[a_i] * beta * np.abs(position[a_i, :] - best_position))
elif f_i == f_g:
e = 1e-20
k = np.random.uniform(-1, 1, 1)
new_p = (inertia_weight[a_i] * position[a_i, :] +
k * ((np.abs(position[a_i, :] - worst_positon)) / (f_i - f_w + e)))
# 越界处理
new_p = np.clip(new_p, search_num_l, search_num_u)
new_f = fitness_function(new_p)
# if new_f.all() < fitness[a_i].all():
position[a_i] = new_p
fitness[a_i] = new_f
if t == 0:
convergence_curve[t] = np.min(fitness)
else:
convergence_curve[t] = min(np.min(fitness), convergence_curve[t-1])
if flag_p == 1:
print(t + 1, " / ", max_iterations)
return convergence_curve
# my_convergence_fit = my_sparrow_search_optimization(population_size,
# max_iterations,
# -search_range,
# search_range,
# Dn,
# fitness_function)
#
# iterations = np.linspace(0, max_iterations-1, len(my_convergence_fit), dtype=int)
# plt.yscale('log')
# plt.xlabel('iterations')
# plt.ylabel('fitness')
# plt.title('sparrow search algorithm')
# plt.plot(my_convergence_fit)
# plt.show()